SciTePress - Affiliation Profile

Cemagref, France

Context (please use complete words only)

Field

Sort by

6 Papers

DISPERSION EFFECT ON GENERALISATION ERROR IN CLASSIFICATION - Experimental Proof and Practical Algorithm

Topics: Neural Networks; Machine Learning; Pattern Recognition

Authors: Benoît Gandar ¹ , Gaëlle Loosli ² and Guillaume Deffuant ³

Affiliations: ¹ Clermont Université, Université Blaise Pascal and Cemagref de Clermont-Ferrand, France , ² Clermont Université, Université Blaise Pascal and CNRS, France , ³ Cemagref de Clermont-Ferrand, France

Digital Object Identifier: 10.5220/0003293007030706

Proceedings of ICAART 2011, ISBN: 978-989-8425-40-9

Quick Abstract

Recent theoretical work proposes criteria of dispersion to generate learning points. The aim of this paper is to convince the reader, with experimental proofs, that dispersion is a good criterion in practice for generating learning points for classification problems. Problem of generating learning points consists then in generating points with the lowest dispersion. As a consequence, we present low dispersion algorithms existing in the literature, analyze them and propose a new algorithm.

Related Ontology Subjects/Areas/Topics (22)

Artificial Intelligence	Knowledge Discovery and Information Retrieval	Pattern Recognition
Biomedical Engineering	Knowledge-Based Systems	Physiological Computing Systems
Biomedical Signal Processing	Machine Learning	Sensor Networks
Computational Intelligence	Methodologies and Methods	Signal Processing
Data Manipulation	Neural Networks	Soft Computing
Evolutionary Computing	Neurocomputing	Symbolic Systems
Health Engineering and Technology Applications	Neurotechnology, Electronics and Informatics	Theory and Methods
Human-Computer Interaction

DOCUMENT CLASSIFICATION - Combining Structure and Content

Topics: Knowledge Management; Data Mining

Authors: Samaneh Chagheri ¹ , Sylvie Calabretto ¹ , Catherine Roussey ² and Cyril Dumoulin ³

Affiliations: ¹ Université de Lyon, France , ² Cemagref, France , ³ , France

Digital Object Identifier: 10.5220/0003505100950100

Proceedings of ICEIS 2011, ISBN: 978-989-8425-53-9

Quick Abstract

Technical documentation such as user manual and manufacturing document is now an important part of the industrial production. Indeed, without such documents, the products can neither be manufactured nor used according to their complexity. Therefore, the increasing volume of such documents stored in the electronic format, needs an automatic classification system in order to categorize them in pre-defined classes and to retrieve the information quickly. On the other hand, these documents are strongly structured and contain the elements like tables and schemas. However, the traditional document classification typically classifies the documents considering the document text and ignoring its structural elements. In this paper, we propose a method which makes use of structural elements to create the document feature vector for classification. A feature in this vector is a combination of the term and the structure. The document structure is represented by the tags of the XML document. The SVM algorithm has been used as learning and classifying algorithm.

Related Ontology Subjects/Areas/Topics (15)

Artificial Intelligence	Enterprise Information Systems	Sensor Networks
Biomedical Engineering	Health Information Systems	Signal Processing
Data Engineering	Information Systems Analysis and Specification	Society, e-Business and e-Government
Data Mining	Knowledge Management	Soft Computing
Databases and Information Systems Integration	Ontologies and the Semantic Web	Web Information Systems and Technologies

INNER AND OUTER CAPTURE BASIN APPROXIMATION WITH SUPPORT VECTOR MACHINES

Topics: Machine Learning in Control Applications; Adaptive Signal Processing and Control; Optimization Algorithms

Authors: Laetitia Chapel ¹ and Guillaume Deffuant ²

Affiliations: ¹ Université de Bretagne Sud, France , ² Cemagref, France

Digital Object Identifier: 10.5220/0003536000340040

Proceedings of ICINCO 2011, ISBN: 978-989-8425-74-4

Quick Abstract

We propose a new approach to solve target hitting problems, that iteratively approximates capture basins at successive times, using a machine learning algorithm trained on points of a grid with boolean labels. We consider two variants of the approximation (from inside and from outside), and we state the conditions on the machine learning procedure that guarantee the convergence of the approximations towards the actual capture basin when the resolution of the grid decreases to 0. Moreover, we define a control procedure which uses the set of capture basin approximations to drive a point into the target. When using the inner approximation, the procedure guarantees to hit the target, and when the resolution of the grid tends to 0, the controller tends to the optimal one (minimizing time to hit the target). We use Support Vector Machines as a particular learning method, because they provide parsimonious approximations, from which one can derive fast and efficient controllers. We illustrate the method on two simple examples, Zermelo and car on the hill problems.

Related Ontology Subjects/Areas/Topics (6)

Adaptive Signal Processing and Control		Intelligent Control Systems and Optimization		Optimization Algorithms
Informatics in Control, Automation and Robotics		Machine Learning in Control Applications		Signal Processing, Sensors, Systems Modeling and Control

AUTONOMOUS MANEUVERS OF A FARM VEHICLE WITH A TRAILED IMPLEMENT IN HEADLAND

Topics: Mobile Robots and Autonomous Systems; Vehicle Control Applications

Authors: Christophe Cariou ¹ , Roland Lenain ¹ , Michel Berducat ¹ and Benoit Thuilot ²

Affiliations: ¹ Cemagref, France , ² Clermont Université, Université Blaise Pascal, France

Digital Object Identifier: 10.5220/0002875501090114

Proceedings of ICINCO 2010, ISBN: 978-989-8425-01-0

Quick Abstract

This paper addresses the problem of path generation and motion control for the autonomous maneuvers of a farm vehicle with a trailed implement in headland. A reverse turn planner is firstly investigated, based on primitives connected together. Then, both steering and speed control algorithms are considered. When the system is driving forward, the control algorithms are based on a kinematic model extended with additional sliding parameters and on model predictive control approaches. When the system is driving backward, two different steering controllers are proposed and compared. Real world experiments have been carried out with an experimental trailer hitched to a mobile robot. At the end of each row, the reverse turn is automatically generated to connect the next reference track, and the maneuvers are autonomously performed. Reported experiments demonstrate the capabilities of the proposed algorithms.

Related Ontology Subjects/Areas/Topics (4)

Informatics in Control, Automation and Robotics		Robotics and Automation		Vehicle Control Applications
Mobile Robots and Autonomous Systems

COMPARISON OF ANFIS AND ORDINARY KRIGING TO ASSESS HYDRAULIC HEAD DISTRIBUTION - The Orgeval Case Study

Topics: Neural Network Software and Applications; Neuro-Fuzzy Systems and Hybrid Systems

Authors: Bedri Kurtulus ¹ , Nicolas Flipo ¹ , Patrick Goblet ¹ , Guillaume Vilain ² , Julien Tournebize ³ and Gaëlle Tallec ³

Affiliations: ¹ Mines ParisTech, UMR Sisyphe, France , ² Université P. et M. Curie & CNRS, UMR Sisyphe, France , ³ Cemagref, France

Digital Object Identifier: 10.5220/0002319903710378

Proceedings of ICNC 2009/IJCCI, ISBN: 978-989-674-014-6

Quick Abstract

In this study, two methods are evaluated for assessing hydraulic head distribution in an aquifer unit. These methods consist in Ordinary Kriging (OK) and Adaptive Neuro Fuzzy based Inference System (ANFIS). Both methods are applied on the same case study: a part of the agricultural basin of the Orgeval located 70 km east of Paris, France. 68 samples were used to predict hydraulic head distribution on a 100 m square - grid. Cartesian coordinates of the samples were used as inputs of the ANFIS, which gives encouraging result. Both simulations have realistic pattern (R2 > 0.97) even if OK performs slightly better than ANFIS at sampling site. Simulated hydraulic head distributions present discrepancies because the two methods capture different patterns. Combined use of the two approaches allow for improving the sampling location of the observation network.

Related Ontology Subjects/Areas/Topics (21)

Artificial Intelligence	Health Engineering and Technology Applications	Neurotechnology, Electronics and Informatics
Artificial Intelligence and Decision Support Systems	Human-Computer Interaction	Pattern Recognition
Biomedical Engineering	Methodologies and Methods	Physiological Computing Systems
Biomedical Signal Processing	Neural Network Software and Applications	Sensor Networks
Computational Intelligence	Neural Networks	Signal Processing
Enterprise Information Systems	Neurocomputing	Soft Computing
Fuzzy Systems	Neuro-Fuzzy Systems	Theory and Methods

ON THE USE OF OPTIMIZATION METHODS FOR THE MINIMIZATION OF FERTILIZER APPLICATION ERROR WITH CENTRIFUGAL SPREADERS

Topics: Mechatronic Systems; Optimization Algorithms

Authors: Teddy Virin ¹ , Jonas Koko ² , Emmanuel Piron ¹ and Philippe Martinet ³

Affiliations: ¹ Cemagref, France , ² LIMOS, Université de Blaise Pascal, France , ³ LASMEA, Université de Blaise Pas, France

Digital Object Identifier: 10.5220/0001212501240129

Proceedings of ICINCO 2006, ISBN: 978-972-8865-59-7

Quick Abstract

Fertilizer application is one of the most important operations in agricultural production. Thanks to their low cost and robustness, centrifugal spreaders are widely used to carry out this task. However, when distances between successive paths followed by the tractor in the field are not constant, application errors occur. These ones generally cause waters pollution and economic issues. In this paper, to limit harmful environmental effects and disastrous drop in production due to centrifugal spreading, we propose an approach based on optimization techniques to improve the fertilization quality. An optimization criterion relying on a spatial distribution model, obtained in previous works, is considered. To compute optimal parameters which should be used as reference variables for the control of the spreader in the future, mechanical constraints are introduced. Faced with a large scale problem, we use an augmented lagrangian algorithm combined with a l-bfgs technique. Simulations results show low application error values comparing to fertilization inaccuracies found without optimization.

Related Ontology Subjects/Areas/Topics (4)

Informatics in Control, Automation and Robotics		Mechatronic Systems		Optimization Algorithms
Intelligent Control Systems and Optimization

Topic(s):

Title:
ISBN:
Year:
Acronym:
Subject:

Title:
Author:
Affiliation:
Subject:

Name:
Affiliation:
Country:
Conference:
Subject:

Name:
Country:
Subject:

Affiliation

Filter listed results, by